Have you ever noticed how addictive it is to search for Q&As? You start with a specific question and soon realize a world of interesting related topics. You could be interested in growing tomatoes, painting furniture, wildlife, or child-rearing concerns. Perhaps you’d like to know the uses of baking soda, or how to get rid of acne. Search engines are not great for exploring Q&As on a topic… what you really need is a "browsing engine"!

Yahoo! Quest offers a new way of browsing Q&A repositories in an intuitive manner, drilling down to the topics that interest you the most. It gives you a quick picture of what’s in the repository, and the discussions related to your area of interest. It does this by analyzing all the questions and "tagging them" automatically with the most important terms.

Want to learn more? Try it for yourself. Choose a topic of interest, type it into the Yahoo! Quest search box, and start browsing for Q&As in your areas of interest.

Yahoo! Quest is currently a prototype and is not integrated into Yahoo! Answers. Instead, we used a snapshot of Yahoo! Answers U.S. from November 2007 containing 8 million Q&As (this data is available for research through the Yahoo! Webscope program). There are several technical challenges in building such a demo. The first one is to select the right "lexical units" of the collection in order to produce meaningful browsing suggestions. The next challenge is to develop interesting list suggestions, on the fly, for whatever query the user may submit. Lastly, we had to invent an interface that would allow users to interact with the suggestions and the results, and enable a natural browsing experience.

We would like to thank some of the people who made this demo possible. First and foremost, we want to thank our ex-colleague Mihai Surdeanu for pushing semantic role labeling beyond the call of duty. Next we’d like to thank Sebastiano Vigna (from the University of Milano, co-creator of MG4J and Archive4J) and Giuseppe Attardi (from the University of Pisa, creator of DeSR) for their wonderful open source code and their life-saving visits. We’d also like to thank Massimiliano Ciaramita, Paolo Boldi (the other creator of MG4J) and Nathalie Vento for their help and encouragement.

Concept, design, research and back-end development was done at Yahoo! Research Barcelona by Jordi Atserias and Hugo Zaragoza. Front-end design and development was done at Yahoo! Research Silicon Valley by Pras Sarkar and Tom Gulik. And special thanks to many of the other Yahoos who collaborated in the creation of this demo, especially Ricardo Baeza-Yates, Roi Blanco, Janet George, Tom Maher and Peter Mika.